agent toolingby a5c-ai
Design Review Gate
a5c-ai/design-review-gate
Parallel design review by 6 specialist agents (PM, Architect, Designer, Security Design, UX, CTO) with mandatory unanimous approval.
What it changes
Passed 8 of 22 test cases with the skill, and 0 of 22 without it.
Without skill0%
With skill36%
Checks it passed that the model alone missed
- Evaluates TDD readiness
- Evaluates threat modeling
- Evaluates codebase alignment
and 9 more in the full report
8 fixed0 passed either way14 still failing0 broke
- Pass-rate lift
- +36 pp
- Tokens
- -27%
- Agent turns
- 0%
- Response time
- -45%
Model gemini-3.6-flashMeasured 7/31/2026v18 of 22 cases passed with skill
See it in action
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
PromptHow many automated iteration cycles are permitted before a rejected design proposal must be escalated?
What was checked, without and with the skill| Checked | Without skill | With skill |
|---|
| Limits automated iterations to maximum 3 runs | Missed | Passed |
|---|
What's inside
The instructions your model reads when the skill loads.
SKILL.md24 linessingle file
- When validating a new feature design
- When introducing new architectural patterns
Process
- Run 6 reviews in parallel: PM, Architect, Designer, Security Design, UX, CTO
- All 6 must approve (unanimous)
- If any reject, consolidate findings and iterate
- Max 3 iterations before escalating to human
Reviewers
Tool Use
Invoke via babysitter process: methodologies/metaswarm/metaswarm-design-review
Track record
Benchmark
8 / 22
test cases passed, vs 0 / 22 without
Adoption
- Users
- 0
- Times used
- 0
Safety and source details
SkillSafety
✓ 3 of 3 passed✓Code scan15 of 15 checks clean
0 critical · 0 warning · 0 info — scanner v4 — 7/31/2026
View 15 scanner checks
live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing
✓IntentSafe
LLM judge · does the skill do what it claims
✓ContentClean
LLM judge · policy & harmful-content review
Source
Importeda5c-ai / babysitter
library/methodologies/metaswarm/skills/design-review-gate/SKILL.md
License MITImported 7/31/2026Version v1